It is possible to protect the privacy of data stored in the cloud and to have more control over who has access to what by using attribute-based encryption. In attribute-based encryption, the key is derived from the data itself, which explains why this is the case. When visible qualities are included in encrypted communication to help users identify important data on vast networks, attribute security has been violated. These obvious qualities are presented to aid users in recognizing important information. This work proposes anonymous key-policy attribute-based encryption, which enables fine-grained data retrieval while maintaining attribute privacy. The goal of this plugin is to provide effective attribute-based access control along with an authorized search approach. By using the secret key that the owners of the data have provided to the users, they can build the trapdoor using the EACAS data. After that, they can carry out searches based on the access constraints to obtain the required information. The design of trapdoors and the deployment of cryptographic protocols both make advantage of a synthetic characteristic that has no syntactic meaning in order to offer attributebased searching on exported encoded data. This is done to make the information accessible. Upon knowledge which has been outsourced, granular access control can be applied in a variety of ways. Establishing the search criteria that data consumers will use to find relevant details based on protected attributes is one strategy that data owners can utilize. This makes it easier for data consumers to find content that is pertinent to their requirements. This paper demonstrate that compared to the most advanced methods now available, EACAS uses significantly less time and storage space for data processing and storage.

Detecting anomalies in videos poses a significant challenge due to the unbounded, infrequent, ambiguous, and irregular nature of abnormal events in real-world scenes. Recently, transformers have shown remarkable modeling capabilities for sequential data. As a result, we endeavor to leverage transformers for video anomaly detection. This paper presents a novel prediction-based method for video anomaly detection called CNNViT by integrating the architectural elements of Convolutional Neural Network (CNN) and Vision Transformer (ViT). The purpose of this fusion is to effectively capture enhanced spatial-temporal information and global features. The effectiveness of the proposed method is evaluated on UCSD Ped2 and CUHK Avenue benchmark datasets. Experimental results demonstrate that the proposed method attains considerably superior performance compared to state-of-the-art techniques.

In our day-to-day lives, social media is becoming a pervasive platform that is misleading society with fake information. In addition to the rise in utilization of social media sites like Facebook and Twitter, news is quickly circulated among the millions of users of Twitter and Facebook within a relatively short period of time. The aim is to provide the user with the ability to classify the news as fake or real and also check the authenticity of the website providing the news. Fake news may lead to very serious conflicts within the nation and disrupt communal harmony. In this study, a machine learning based approach for Fake News Detection and Severity Determination(FNDSD) is proposed for the same. In the proposed FNDSD model, the severity of the impact caused by the fake news is determined by proposing two methods : (i) A clustering based method named as Severity Determination using Clustering(SDC) and (ii) Severity Determination by Finding Category(SDFC). A degree of severity on a scale of 0 to 1 is taken as a measure, and fake news is assigned low, medium, or high severity accordingly. In the proposed SDC method, fake news is said to have more severity if its degree of severity is greater than 0.5, and in the category-based severity method, fake news degree of severity is determined using its category. Better fake news classification results are observed in case of ISOT dataset when compared to Liar and FA-KES datasets.

The prospective utilizations of medical imaging data for the dependable identification and authentication of individuals have garnered considerable attention in both security and healthcare realms. This significance becomes particularly pronounced in scenarios involving disasters, where conventional means of human identification prove ineffective. Amid such circumstances, Chest X-rays emerge as a valuable resource, capturing the distinct anatomical features of an individual's rib cage, lungs, and heart. These features, inherently unique, stand as steadfast markers for identification, even when the human body is compromised. Given these premises, the integration of Artificial Intelligence techniques, particularly deep learning methodologies, proves exceptionally beneficial in discerning distinctive identifiers within chest X-ray images. This study capitalizes on the potential inherent in a deep learning paradigm, specifically the Siamese network and Triplet loss, presenting an innovative approach to person identification through chest X-ray images. While prior research in this domain has predominantly relied on Convolutional Neural Networks (CNNs) and conventional machine learning methods, our work introduces a distinctive framework by advocating the utilization of Siamese and Triplet networks for this pivotal task. This contribution expands the existing body of knowledge by pioneering a novel methodology that harnesses the Siamese and Triplet loss deep learning architecture to enhance the processing of Chest X-ray images. The implications of our work reverberate significantly in the domains of security and healthcare, where the refinement of dependable and precise identification techniques holds paramount importance, particularly in emergency scenarios.

Single Sign-On (SSO) systems have become popular for simplifying the login process by allowing users to authenticate with a single identity provider (IDP). However, the widespread use of these systems raises concerns about user privacy, as IDPs like Google or Facebook can gather extensive information about user behavior on the web. This poses a significant challenge for privacy-conscious users who wish to limit the disclosure of their web access behavior to third-party organizations. In this paper, we present a comprehensive and large-scale practical study focused on the OpenID Connect protocol, a widely used SSO protocol. Our analysis delves into the various stages of the protocol's operation, highlighting its security flaws and vulnerabilities. Additionally, we conduct a systematic examination of the privacy implications associated with user access to SSO systems. We provide a detailed account of how easily user information can be accessed, shedding light on the potential risks.

The primary objective of this research endeavor was to comprehensively examine the influence of artificial intelligence (AI) applications, encompassing expert systems, neural networks, and genetic algorithms, on the qualitative attributes of data within commercial banks listed on the Amman Stock Exchange. In order to fulfill the goals of this study, a dualpronged methodology was adopted. The first approach was the descriptive-analytical method, which entailed the development of a questionnaire to gather primary data pertinent to the independent variables associated with AI applications. The second approach, the applied method, involved the assessment of the dependent variable, denoting the qualitative characteristics of data, utilizing the financial statements of commercial banks listed on the Amman Stock Exchange during the period spanning from 2015 to 2022. The analysis of data was conducted employing appropriate statistical techniques, and the formulated hypotheses underwent rigorous testing via multiple regression analysis. The findings derived from this study unveiled a noteworthy, statistically significant positive impact of varying magnitudes attributed to AI applications on the qualitative characteristics of data within the commercial banks listed on the Amman Stock Exchange. Specifically, the expert systems variable emerged as the foremost influencer, showcasing the highest strength of impact on the attainment of qualitative data characteristics. It was closely followed by the genetic algorithms variable, with the neural networks variable trailing in its impact. These results serve as a clarion call to the commercial banks in Jordan, urging them to bolster their utilization of AI domains and to encourage their effective application. This entails assigning a more prominent role to information technology (IT) experts, smart applications, and genetic algorithms. Such measures are aimed at ensuring the efficient and cost-effective utilization of AI domains, with a view to enhancing the quality of produced data's qualitative characteristics, promoting transparency, instilling trust among investors, and optimizing the efficiency of banking operations in a manner that underpins profit sustainability.

Continuous innovation in virtual reality (VR) technology has increased the demand for digital learning experiences and reduced time and costs while implementing sophisticated design projects in digital marketing education. VR learning hardware like the CAVE room, which provides computer-generated 3D prototypes, has increased in education due to its advantage of Visualizing, analyzing, optimizing, and testing it before the final product is produced. Following a literature review, this research explored the students' learning outcomes based on a spatial understanding of the complex digital marketing course design, collaboration, engagement, and problem-solving skills as VR technology is integrated into higher education. The study utilized mixed-method research, quasi-experiments, questionnaires, and experiments based on project-based learning on 94 Al-Zaytoonah University of Jordan students. The qualitative and quantitative results were interpreted separately before being used to inform the decision and conclusion. The study found a significant impact of VR on education and learners' outcomes, as indicated by the standard deviation of the VR approach post-test assessment and the high-quality presentation skills during class presentations.

In the landscape of healthcare, diabetes remains an urgent concern, demanding effective detection methods due to its widespread impact. In response, our research takes up the mantle, harnessing cutting-edge machine learning algorithms to advance diabetes prediction. Our core methodology revolves around pioneering deep learning algorithms, which unravel intricate patterns within complex medical datasets. Through rigorous evaluation, our study demonstrates the proficiency of this approach in accurately distinguishing between positive and negative diabetes cases. Notably, our investigations reveal that Deep Learning - Scenario 1 emerges as the most promising methodology, achieving an accuracy of 84.24%, surpassing the accuracy attained by Logistic Regression (82.46%), Deep Learning - Scenario 2 (69.48%), and Deep Learning - Scenario 3 (80.21%). This remarkable outcome underscores the potential of Deep Learning - Scenario 1 in enhancing diabetes prediction accuracy. The fusion of innovative methodology and empirical outcomes marks a significant leap forward in diabetes management, highlighting the transformative power of machine learning applications. Through this synergy of inventive techniques and promising results, our work resonates with the evolving realm of diabetes research and healthcare practices.

Over the past decade, Intelligent Vehicles (IVs) have revolutionized operations in various sectors of which the healthcare sector is not an exception. The integration of IVs with the healthcare systems tackles various issues and challenges and ultimately leads to improved care of patients, remote monitoring of patients, and quick and efficient response mechanisms for emergencies. The integration of IVs with trending technologies in the healthcare sector highly nurtures the performance of the healthcare systems. In this article, we present a comprehensive analysis of the integration of IVs with healthcare systems and highlighting the applications of IVs in the healthcare sector. The integration of IVs with healthcare system pose various issues and challenges that are presented in this work. Lastly, the paper presents future research directions which will be helpful for the researchers to explore the relevant domain.

Air pollution appears in the form of outdoor air quality and indoor air quality (IAQ). Particulate Matters (PM_2.5and PM₁₀) and CO₂, among many air pollutants, are responsible for worsening IAQ. IAQ has been linked to lung illnesses such as asthma, chronic obstructive pulmonary disease (COPD), and lung cancer. This linkage is related to how much any individual gets exposed to poor IAQ, individuals' daily indoor activities, and related health risks. The study presented in this paper focused on the analysis and identification of the relationship between IAQ, daily indoor activities and health concerns and their effect on breathing issues in five asthma patients in Bradford, UK. IAQ data were collected using an IoT-enabled, low-cost device capable of monitoring CO₂, PM_2.5, PM₁₀, relative humidity, and temperature. In addition, indoor daily activities and breathing issues that relate to asthma severity are also recorded using a digital platform. These recorded data are statistically analysed using correlation and measure of central tendency. Results indicate a strong positive correlation between breathing issues and indoor measured pollutants (0.72, 0.62 and 0.63 correlation coefficient with CO₂, PM₁₀and PM_2.5respectively). IAQ is strongly correlated with indoor activity in terms of window opening hours and breathing issues as asthma symptoms. It was observed that when the mean readings of PM_2.5, PM₁₀and CO₂are higher (22μg/m3, 24μg/m3 and 700 ppm respectively - in this study), the asthma patients reported experiencing worsening symptoms. This study highlights the importance of managing IAQ and window-opening habits to potentially mitigate asthma symptoms and improve patient care.

Batteries are critical in the creation and operation of smart transportation systems such as electric vehicles (EVs), as they provide efficiency, dependability, and endurance. Accurate State of Health Estimation (SoH) estimation guarantees that the battery can properly store and transfer energy, maximizing the vehicle's range and performance. By extending battery life and lowering the need for premature replacements, battery SoH estimation contributes to sustainable mobility. This minimizes the environmental impact of battery manufacturing and disposal. Present study focused on four specific batteries tested at NASA's Prognostic Centre of Excellence. The study utilized Artificial Neural Networks (ANN) and Random Forest (RF), to estimate discharge capacity using training, testing, and validated with ten-fold cross-validation. In the current analysis, three accepted metrics were used to evaluate the effectiveness of models. In terms of prediction accuracy, the findings showed that RF consistently beat ANN, with lower Root Mean Square Error (RMSE) and Mean Absolute Error (MAE) values and higher Coefficient of Determination (R²).

The detrimental effects of air pollutants on human health have prompted increasing concerns regarding indoor air quality (IAQ). The emergence of digital health interventions and citizen science initiatives has provided new avenues for raising awareness, improving IAQ and promoting behavioural changes. The Technology Acceptance Model (TAM) offers a theoretical framework to understand user acceptance and adoption of IAQ technology. This paper presents a case study using the COM-B model and Internet of Things (IoT) technology to design a human-centred digital visualisation platform, leading to behavioural changes and improved IAQ. The study also investigates users' acceptance and adoption of the technology, focusing on their experiences, expectations, and the impact on IAQ. Integrating IAQ sensing, digital health-related interventions, citizen science, and the TAM model offers opportunities to address IAQ challenges, enhance public health, and foster sustainable indoor environments. The analytical results show that factors such as human behaviour, indoor activities, and awareness play crucial roles in shaping IAQ.

In recent years, unmanned aerial vehicles (UAVs) have gained widespread application due to their advantages such as high durability, low cost, ease of implementation, and flexibility. However, traditional UAV applications often focus on a singular task, requiring redesign and retraining when there are changes in scenarios and objectives. To address this issue, we propose a construction approach for a general UAVs task framework based on Large language models (LLMs), enabling adaptive execution of various UAV tasks in a zero-shot manner. In our methodology, we initiate the extraction of capabilities from black-box LLMs serving as victims, employing a mimicry attack to transfer these abilities into the mimic model CodeT5. Subsequently, fine-tuning is performed using UAV domain data, ensuring that the imitation model is specialized in generating UAV control code. To further compress the model for deployment on UAVs, we employ knowledge distillation techniques, utilizing the fine-tuned imitation model as a teacher to train a high-precision lightweight student model. This generative approach facilitates on-the-edge task execution and auxiliary decision-making. During the execution phase, users can achieve real-time online interaction with UAVs through natural language commands.

The usage of mobile devices and programs that can be downloaded onto such devices is becoming more common in day-to-day life. Mobile applications need a significant amount of resources to function properly. On the other hand, computers that need less resources to function properly are considered to be cleverer. By shifting computation and/or data from mobile devices to the cloud in order to free up space on mobile devices, mobile cloud computing (MCC) presents a solution to the problem of limited storage capacity on mobile devices. MCC is an abbreviation for mobile computing in the cloud. MCC has discovered a solution to this problem, which involves shifting the processing and/or data from mobile devices to the cloud. The cloudlet is a helpful addition to the client-cloud hierarchy, particularly in light of the increasing interconnectedness of cloud computing and mobile networking. Within the scope of this work, we have given you with a comprehensive analysis of cloudlet-based mobile computing science for your perusal and consideration. Following that, we will discuss two problems that end users have while using cloudlets. In this last section, we take a look at the current state of this business as well as the opportunities and challenges that lie ahead for it in the foreseeable future. This is especially true for real-time applications such as augmented reality, such as the ones that were discussed before in this paragraph. As a direct consequence of this, the cloud itself is required to be sent to the location of the mobile user in the form of cloudlets. In contrast to the traditional idea, which necessitated the use of specialist equipment, the creation of cloudlets is now possible without the use of any such apparatus. Because of this, the creation of cloudlets moves forward much more quickly.

Image watermarking, a technique of documents protection has been devoted to resolve various issues related to digital content security and copyright authentication. In this paper, a model of content watermarking is proposed. It is built based on DWT, DCT and SVD techniques. Various experiments have been processed. The evaluation results show an effective performance of the proposed model in regard of similar available works.

This study aims to discern prominent current and future trends within the realm of accounting and corporate sustainability, as extracted from the Web of Science database, with a focus on publications spanning the period from 2010 to 2023. Some research has specifically focused on the digital business environment. Building on the resource-based view theory, this paper investigates the evolution of auditing and digital accounting research in the digital business environment through bibliometric analysis. Utilizing the VOSviewer tool, the study analyzes data concerning both keyword relationships and their geographical distribution. A total of 700 publications sourced from the Web of Science have been categorized. Additionally, the findings highlight the transformative nature of auditing and digital accounting in response to the rapid digitization of business operations. The emergence of key clusters, which focus on digital transformation and explore the relationship between information technology and corporate management, indicates the evolving dimensions of research in these fields. Both the practical and theoretical implications have been discussed.

This assignment poses a challenging but essential problem for the rapidly developing discipline of natural language processing (NLP), which tries to classify information that is contested within texts that cover a variety of subject domains. The assignment in question seeks to classify information that is contested within texts. Enhanced Deep Inference for Text Segmentation (EDITS) is the name of the model that was used in this research project to demonstrate how deep learning can be integrated with more conventional approaches to machine learning such as Random Forest, Support Vector Machine, and Naive Bayes. The purpose of this research was to demonstrate the usefulness of deep learning in practice. The naive Bayes algorithm, support vector machines, and random forests are some of the approaches that fall under this category. In order to extract features and semantic information from the writings that are a part of the Quantitative Argumentation Dataset (QuAD), which is made of writings on a wide variety of topics, we use methods such as term frequency-inverse document frequency (TF-IDF) analysis. This allows us to extract information from the writings that is relevant to our research. This allows us to extract information from the texts with a greater degree of precision. Because of this, we are able to glean information from the sentences in a more accurate and exact manner. In this study, we compare a deep learning model called EDITS, which was developed specifically for lectures, to models that are more traditional. EDITS was created with the intention of improving speech recognition. This research expands our understanding of argumentation in a wide variety of contexts and demonstrates how cross-topic dispute may be classified via EDITS in real-world scenarios.

In the aftermath of the Covid-19 pandemic, wearing face masks has become a widespread precautionary measure, presenting new challenges for surveillance systems. To tackle this challenge, we introduce an innovative surveillance system that applies deep learning methodologies to detect anomalies in the behavior of masked persons. We adopted a uniquely designed deep neural network for spotting real-time video stream abnormalities and implemented efficient techniques, using the Dlib library and other image processing methods. The neural network underwent training on a diverse dataset encompassing both normal and abnormal activities of masked individuals, enhancing its capability to discern suspicious behaviors effectively. Through extensive experiments conducted on publicly available datasets, our system demonstrated exceptional performance in the anomaly detection of masked individuals, achieving an accuracy rate of 90%.

With the continuous advancement of technology, the risk of cyberattacks has become increasingly prominent. Recent advancements have demonstrated that modern, powerful computers such as quantum computers could potentially break even robust encryption algorithms like RSA. Although achieving this level of computational power currently requires a tremendous amount of time and resources, technological advancements may soon make hacking more feasible. As people become increasingly reliant on technology for daily operations, cyberattacks pose a more significant threat than ever before. The Internet of Things (IoT) is one technology that has rapidly gained popularity, with a growing demand for devices such as personal home cameras, nanny cameras, and other consumer-grade IoT-enabled video surveillance devices. To enhance the security of message transmission in IoT-based smart surveillance systems, we propose a more secure and efficient approach utilizing Zenoh, one of the newest and most performant publish/subscribe message exchange technologies. We have implemented various Post-Quantum Cryptography (PQC) algorithms, namely CRYSTALS-KYBER and NTRU, to encrypt the data. Following the implementation, we evaluated the performance of each algorithm to determine the most suitable option for IoT implementation.

Detecting and diagnosing dermatological diseases is of paramount importance in ensuring early and accurate medical interventions, thus improving patient outcomes. In this context, our study presents a pioneering approach to derma disease detection utilizing machine learning algorithms. We introduce a novel methodology that leverages Deep Learning - Keras, Deep Learning - CNN, and Deep Learning - Transfer Learning algorithms to classify skin conditions. The results demonstrate the efficacy of our approach, achieving accuracy percentages of 56.79%, 55.16%, and 58.15% respectively, across different models. Among these, the Transfer Learning algorithm emerges as the most accurate, underscoring its potential for revolutionizing dermatological diagnostics and offering a promising avenue for more precise and efficient disease detection.

Recently, many businesses and industries involved sustainability in their strategies. The demand for sustainable development in business has drawn the attention of the government and academia and emerged as a broadly well-known trend. The aim of this project is to develop an integrated approach of the sustainable business from a theoretical perspective integrated with the current wave of big data analytics. To achieve this, aim an advance methodology is employed for the assessment of business sustainability based on the proposed big data analytics model from consumer perception dimension. In addition, this research will quantify the effect of big data analytics on business sustainability. It also proposes a scale for measuring and monitoring sustainability in the industry benchmarking the experiences of big data analytics usage globally and locally. The method used to achieve the mentioned aim is a qualitative approach. Articles published from 2015 to 2022 will be reviewed to get the findings and results of this review. This study contributes to the advancement of knowledge and literature by offering a comprehensive scale based on big data analytics to measure sustainability in the industries. Moreover, this research project is expected to produce research findings that will enhance our understanding of the importance of data analytics in business sustainability. Many literatures showed the impact of data analytics on business performances, but there are less articles that demonstrated the impact of big data analytics on business sustainability which aimed to be explored by this research project.

Computers and mobile phones are common today and play an important role in society for children and adults, providing many opportunities and advantages. This new application is intended for everyone, including children, and offers an enjoyable learning process. It takes advantage of children's natural ability to identify individuals, objects, scenery, and other visual features in still photos and moving clips. This paper aims to create an educative but engaging environment that stimulates children to learn about and understand various visuals. In addition, the incorporation of deep NN models into the software demonstrates the need to modify software engineering methodologies and practices to cope with evolving technologies. The program can recognize an array of objects, such as animals, cars, common housewares, and natural scenes. The software uses state-of-the-art deep neural network architectures that are readily plugged into modern software stacks for accurate and reliable object detection. Aside from merely identifying objects, it also engages users in the learning process, where they are prompted to recognize and label different objects and parts of the visual content. This method provides a stimulus or motivation for children to learn object names in English because it facilitates the swift and faster growth of their own vocabularies. Therefore, this project is at the crossroads of computer vision, deep learning, and education, as such tools contribute to the enhancement of children's language knowledge and object identification abilities in mobile application development.

Massive news streams and several Online Social Networks (OSN), such as Twitter today necessitate automated analysis. But the majority of them don't offer sentiment analysis. The primary issue with news sentiment analysis is the disparity in reader and writer opinions toward the news text. News might be positive or negative, but it must be presented as simple facts in impartial language. The issue of sentiment analysis in OSN is still a real challenge because the most recent news has an impact on people's lives every day, despite the fact that there exist programs for sentiment analysis of news. Immediately following the report of a potential Russian invasion of Ukraine, OSN users from all over the world began to flood Twitter with tweets. We can better comprehend public opinion on various events before and during the war by analyzing these opinions. For sentiment analysis, 10009 distinct tweets from January 2022 to the third week of July 2022 were examined. We investigated the problem of sentiment analysis for Ukrainian and Russian news in this paper, created a corpus of Ukrainian and Russian news, and annotated each text with one of three categories: positive, negative, or neutral. Overall, 37.06 percent of tweets are positive, 22.90 percent are negative, and 39.5 percent are neutral.

This research examines how cloud and intelligent edge computing are affecting smart city physical environments. This shows how these technologies might transform urban life. Cloud and intelligent edge computing may improve urban life and sustainability. These applications are useful across sectors. Integrating centralized cloud resources with decentralized edge devices makes these applications operational. This technology may make cities more livable and environmentally friendly. This article focuses on how recent technological breakthroughs are affecting transportation, healthcare, energy management, and public safety. The combination of edge and cloud computing makes smart cities feasible and accessible. These cities are efficient, adaptive, and environmentally conscious. The convergence's environment will spur technology and urban innovation, which will change how people live in cities in the future. Additionally, these qualities will create an environment that encourages urban planning and technological innovation. This research examines the benefits of cloud and intelligent edge computing for various enterprises. These technology advances are affecting several industries, including public safety, energy management, transportation, and healthcare. This article devotes a lot of attention to each of these specific areas. This study shows the necessity of communication protocols and standards and how to use them by offering a step-by-step strategy for data collection, analysis, and interpretation. We will focus on message fundamentals and network operation throughout the study. The results show that smart edge and cloud computing in smart cities increase energy efficiency, sustainability, and mobility.

This research focuses on improving NLP abilities using machine learning to build intelligent conversation agents. The information fetched through the literature review and the approach to the research methodology includes collecting, processing, and testing the models with big NLP datasets. The research's goal is to intensely improve the quality of conversational AI by fine-tuning state-of-the-art NLP models for a conversational context. This research was conducted in Jordan, where they exploited the region's unique features and dialects for enhanced knowledge of the natural language by AI, thereby improving the possible uses of conversational AI globally. The expected outcomes are twofold: enhanced conversational capabilities for conversational AI and a significant research contribution to the knowledge base, especially in customer service and virtual assistance. The impact of the research is far-reaching as improvements in conversational AI are expected to ease cross-sector operations, improve user experience, and lay a foundation for future developments in language-based AI interaction. The research stands at the cutting edge of technological development in Jordan and may serve as a model for other regional Conversational AI innovations. The use of machine learning combined with NLP is not only moving forward the conversation abilities of artificial intelligence systems, but it is also preparing the ground for next-generation advancements to change the nature of human-computer interaction. The realization of the findings can be instrumental in opening up a whole new era of AI whereby it talks, comprehends, and even satisfies the needs of humans through unparalleled eloquence and accuracy.

In contemporary society, the pace of our daily lives has reached such a level of intensity that we find ourselves with limited opportunities to engage with crucial informational resources. The act of reading has transitioned from being obligatory to becoming elective. The preference for podcasts over novels and other written media is driven by the perception that podcasts are more time-efficient. In the current era of technological technology, utilizing this medium demands less cognitive effort compared to reading and offers a more user-friendly experience. Optical Character Recognition (OCR) refers to a technique used to audibly interpret text that has been retrieved from photographs or PDF files. The proposed methodology involves converting PDF documents into audio files, while eliminating extraneous parts such as headers, footers, page numbers, publisher copyright seals, external references, and other similar components. The approach employs word recognition and textto- speech technology to convert PDF files into audio versions. The process involves employing several technologies such as Optical Character Recognition (OCR) and text-to-speech synthesizers, like Google Text-to-Speech and Python Text-to-Speech, to convert written text into audible audio. The aforementioned technique effectively mitigates the generation of said elements as undesirable outputs within the ultimate audio file. The study at hand aims to assist individuals who are visually impaired and face challenges in reading. Podcast versions of these texts are available for auditory consumption. The proposed methodology aims to achieve the goal of creating a smart environment in which individuals utilize advanced systems to enhance their quality of life.

There is a lack of AI-based solutions that can detect Cloud-based Brute-Force attacks. The current methodologies followed for the purposes of protecting Cloud environments from Brute-Force attacks involve the incorporation of DES Keys and One-time Passwords (OTPs), centralizing Cloud resource access as well as applying IDS (Intrusion Detection Systems) on different network levels, as well as applying administrative controls over an environment to protect against said attack. This research aims to find an AI-based solution that can scan through Cloud user logs in order to find anomalies and atypical login/ usage patterns and report them to the corresponding technical administrator. The two major limitations of this research are that it focuses on Microsoft Azure as a main platform for applying this solution; rather than having a comprehensive approach that covers all Cloud service providers, we chose to focus on the most utilized one, especially from an organizational point of view, The second limitation of this study is that there is very little research conducted on AI applications in the context of Brute-Force attacks, which can make it a bit challenging to find relevant resources.

False Data Injection Attacks (FDIA) are among the most significant security threats to critical systems. In this paper, we present a comprehensive benchmarking study to evaluate the performance of four deep learning models for detecting FDIA. The study covers an analysis of three diverse datasets representative of real-world scenarios. Our evaluation leverages a comprehensive set of metrics, including precision, recall, accuracy, F1 score, ROC analysis, and precision-recall plots. The primary objective is to identify the optimal model that exhibits robust performance across various attack scenarios. Subsequently, the selected model undergoes in-depth examination through the application of Principal Component Analysis (PCA) to ascertain the influence of dimensionality reduction on its efficacy. The results of this study enhance the understanding of deep learning capabilities in FDIA detection and lay the foundation for the development of advanced detection and defense mechanisms.

The application of artificial intelligence (AI) in diverse disciplines has manifolded in the past years. More recently, ChatGPT launched by open AI has become quite popular in academics as well as in the corporate world. This paper unfolds the extant of ChatGPT utilized during the critical process of determination of the tariff code internationally and nationally. It is argued that due to critical, complex and sensitive nature of tariff classification, deep learning and machine learning techniques can be closer to the correct determination of HS (Harmonized system) codes compare to ChatGPT. Based on multi-steps exploratory nature of experiment conducted over ChatGPT with various chapters, headings and sub-headings of multiple products, our findings show that auto-classification of HS code through ChatGPT has limited scope and level of correctness. In some cases, the output provided by tool is correct to the extent of only heading level and in other cases, the output suggested is completely categorized in an irrelevant HS code. Therefore, the traders, customs authorities, and customs brokers, should not fully rely on this tool as they can face severe consequences in terms of compliance issues and monetary penalties due to incorrect HS code used in a supply chain for trade purposes.

This study examines the influence of AI )Artificial Intelligence (AI) integration on supply chain risk mitigation (SCRM) in the Middle Eastern manufacturing industry. A survey involving 240 managers from the manufacturing sector was conducted, and the data were analyzed using Structural Equation Modeling with Partial Least Squares (SEM-PLS).This study reveals that integrating AI into supply chain (SC) operations has a significant positive impact on mitigating supply chain risk (SCR). Notably, this study highlights the crucial role of Data-Driven Decision Culture as a mediator in this relationship. AI integration plays a critical role in fostering a culture of data-driven decision-making within enterprises, in addition to immediately assisting in the mitigation of SCR. In order to properly manage and lower SCRs, companies, policymakers, and industry professionals can benefit greatly from the research's findings, which can be used to guide their decisions about AI technology investments and the development of data-driven organizational cultures.

This study examines how machine learning and artificial intelligence (AI) may be used to solve urgent worldwide issues with waste and energy management. It presents a thorough framework for employing AI technology to track and optimize trash production and energy use in residential communities. The framework reduces energy consumption and increases waste management efficiency by using sensors, data processing, predictive modelling, and real-time monitoring to enable data-driven decision-making. It emphasizes how AI has the power to transform conventional methods and usher in a more sustainable future. The study offers a cutting-edge framework for waste and energy management monitoring that is based on machine learning and provides real-time insights and forecasts to help with decision-making. Constant data analysis maximizes the use of resources, resulting in reduced costs, increased operational effectiveness, and advantages for sustainability. Although there are still obstacles to overcome, such as data security and upfront costs, this paradigm is an essential first step towards effective and accountable resource management.

This article investigates the relationship between digital transformation and the performance of the ASE-listed firms in Jordan. To achieve the research objectives, cross-sectional time series analysis ``Panel Data" is utilized. The research sample consists of 43 ASE-listed service firms for 5 years between (2018) and (2022). Content analysis adopted as a research instrument is used to collect study data. The dimensions of digital transformation encompass artificial intelligence, blockchain, cloud-based accounting, big data, and digital applications. The results indicate that digital transformation has no impact on the financial performance of firms represented by return on assets (ROA). Another key result is that a positive relationship is found between digital transformation and market performance represented by the variable ```Tobin's Q". Given the said results and the technological and digital development witnessed by the world, it is evident that this research is applicable and essential to understand the significance of digital transformation in developing the performance of firms, as large firms can only survive and compete and therefore firms must pay more attention to digital transformation and its tools.

The application of Visual Artificial Intelligence (AI) in smart surveillance systems has signifi-cantly evolved, particularly in the domain of crime detection and prevention. This survey aims to com-prehensively review and analyze the implementation of visual AI technologies within a smart surveillance framework, focusing on their role in enhancing security measures for crime detection. Most frameworks found in the literature rely on the use of Convolutional Neural Network for image classification, however, areas of improvements are still open particularly the limited availability of datasets, high computational costs, and effective deployment of these types of applications. The survey investigates various facets, in-cluding but not limited to, object detection, facial recognition, anomaly detection, behavior analysis, and scene understanding within the context of surveillance applications. Furthermore, this survey aims to of-fer insights into the current state-of-the-art methodologies, recent innovations, and prospects in the realm of Visual AI-powered smart surveillance, providing a foundation for further research and development in this evolving field.

This study presents a new approach to predict cryptocurrencies risks based on social media platform indicators. This method can help cryptocurrency investors in their future decision-making. It utilizes advanced techniques such machine learning and natural language processing. Furthermore, a case study on American cryptocurrency investors is provided to help them identify and assess the risks associated with their purchases. Moreover, performance metrics are computed to demonstrate the capabilities of the new approach.

The study of decisions and decision making has long been a focus of research. This article discusses and examines the four models of decision processes, rational, bureaucratic, political and organized anarchy and offers observations about how these models are likely to interact, co exist and can be applied to the organizations. It mainly focuses on the value of the delineation as separate theories to describe the salient features of each model. This helps to facilitate the exploration of the dichotomy, which may be discerned in academic thinking over the last four decades, between the analysis and advocacy of incremental/decremental or comprehensive approaches to budgeting and resource allocation in organizations.

Purpose:Recent years have seen dramatic growth in the number of scholarly scientific works dedicated to the topic of Artificial Intelligence (AI), banking, and finance. With the development of scientometric tools, it is now possible to map, visualize, analyze, and assess scientific activities in many fields of engineering and social sciences including Islamic banking and finance.Design/methodology/approach:Based on data retrieved from the Scopus database and using a qualitative method, this study investigates, evaluates, and identifies significant development, trends, and players in the application of AI in Islamic banking and finance using the Visualization of Similarities Method (VOS) between objects in VOSviewer. From 2006 to 2022, 387 documents were retrieved from the Scopus database.Findings:Results revealed that the top five most active countries in terms of publications are Malaysia (117 documents), Indonesia (89 documents), the United Kingdom (36 documents), the USA (24 documents), and Saudi Arabia and Bahrain (24 documents each).Research limitations/implications:The International Islamic University Malaysia and Universitas Indonesia came out on top of active institutions while the top funding source came from the Ministry of Higher Education, Malaysia.Originality/value:The comprehensive findings and analysis reported in this study serve as a roadmap for future academics to create theory and practice for applying AI in Islamic banking and finance.

With the recent evolution ofIndustry 4.0, integrated devices and sensors started to generate dynamic, extensive, and heterogeneous in Industrial Internet of Things (IIoT) environments. The immense heterogeneous data probably raises the difficulties of data analysis procedures and security risks. As IIoT develops, cyber-attacks are more complex and diverse, making current anomaly detection methods' operations less effective. Intrusion Detection System (IDS) had a major contribution to protecting IoT systems from cyberattacks by constructing complementing classical preventive security measures (e.g., encryption, authentication, and authorization). Therefore, in this paper, we present a Levy Swallow Swarm Optimization with Machine Learning-based Intrusion Detection (LSSOML-ID) technique in IIoTenviornments. The presented LSSOML-ID technique focuses on the identification and classification of intrusions in IIoT environments. To achieve this, the LSSOML-ID technique uses a Z-score normalization algorithm to scale the input data into a uniform format. For feature selection purposes, the LSSO algorithm was used which reduces the number of features involved in the classification process. Finally, the Aquila optimizer (AO) algorithm with an extreme learning machine-based autoencoder (ELM-AE) model is proficiently employed for intrusion classification. To exemplify the significant outcomes of the LSSOML-ID technique, a thoroughset of experiments areconducted. The comprehensive outcomes highlighted the betterment of the LSSOML-ID approach over recent methods.

Artificial intelligence (AI) and other AI-based applications are being integrated into firms's HRM and supply chain management. The COVID-19 caused widespread devastation and showed the fragility of AI in HRM and supply chains. Few attempts have investigated the linkage between AI in HRM and supply chains during and post this horrible epidemic. Our research aims to review these attempts to have a glance into the current status and the future direction of this linkage. To that purpose, a systemic literature review technique has been employed. Several academic databases were searched between January 2020 and November 2023. The findings of our review show that the search on AI in HRM and supply chains during and post the pandemic is still in the infant stage and needs more attention. We discussed COVID-19's effects on AI in HRM and supply chains research, as well as outlined some important directions for future research.

Social media influencers have become important motivators shaping tourist attitudes and behaviors. This study analyzed how exposure to influencer content impacts key outcomes for the destination Jordan. A survey of tourists who visited Jordan in the past 3 years measured their perceptions of influencer credibility, content quality, awareness/interest, trust/engagement, general tourism behavior, and intentions to visit. Results of SEM analysis found significant positive effects of influencer marketing on both dimensions of tourist behavior. Awareness/interest and trust/engagement were most impactful, highlighting influencers' role in sparking early motivation. Content quality additionally predicted visit intentions by informing decisions. Perceived credibility made recommendations more persuasive. Findings provide theoretical validation of how influencers act as digital opinion leaders. Managerial implications suggest destinations should invest in influencer campaigns for reach and inspiration, while ensuring content quality. Overall, influencer marketing demonstrated significant persuasive appeal for potential tourists.

Internet of Things (IoT) networks are prone to cyber attacks because of insecure wireless infrastructures, various kinds of IoT devices, large quantities of data transfer by sensors over the network and resource constraint architecture. Hence, IoT-compatible cybersecurity solutions are needed. One of the typical solutions to find cyber threats in IoT- based networks is an automatic detection system. The new trend in cyber security is Machine Learning (ML)-based analytics intends at extracting security-related data to identify advanced target cyber threats and reduce the overhead of keeping static correlation rules. However, many prevailing solutions for cyber threat detection include numerous problems like high false positive rate (FPR), low scalability, high learning complexity, and poor accuracy. Therefore, this study designs a new Flamingo Search Algorithm with Optimal Deep Learning for Automated Cyber Threat Detection (FSAODL-ACTD) technique in the IoT system. The main objective of the FSAODL-ACTD technique is to automate identification of the cyberthreats in IoT networks. In the FSAODL-ACTD method, a new feature selection approach using FSA is derived. Besides, multilayer bidirectional long short term memory (MBiLSTM) model is utilized for the automated identification of the cyber threats in the IoT environment. Finally, sine cosine algorithm (SCA) is exploited for the optimal hyperparameter adjustment of the MBiLSTM method. The simulation outcomes of the FSAODL-ACTD method are tested on the cyberthreat dataset, and the outcomes ensured the supremacy of the FSAODL-ACTD approach over other current approaches.

The transportation industry as a whole is now undergoing a transition towards more environmentally friendly business practices, and electric vehicles (EVs) are an essential component of this industry-wide shift. An intelligent charging algorithm that makes use of multiprocessing strategies and is based on the Internet of Things (IoT) is now the subject of study as a potential solution to the problems associated with inefficient and wasteful charging of electric vehicles (EVs). The development and evaluation of a smart charging algorithm for electric vehicles using the Internet of Things (IoT) is going to be the focus of this project. This integration makes it possible to create real-time communication between electric cars (EVs), charging stations, and the grid, which is necessary in order to offer charging that is intelligent and adaptable. The growth of the electric vehicle sector will only increase the need of coming up with innovative answers to problems. The purpose of this study is to evaluate the effectiveness of the algorithm and to determine the extent to which it has the potential to revolutionize the process of charging electric cars while also improving grid management, energy efficiency, and the prospects for a sustainable future.

This paper focuses on precise energy consumption prediction through AI-driven models using time series data. Addressing data quality and model selection challenges, we conduct thorough data analysis, rectifying missing values, and normalizing inputs. Leveraging LSTM due to its efficacy in capturing temporal dependencies, we surpass existing limitations. Experimental outcomes validate our approach, accentuating the significance of data analysis, highlighting LSTM's relevance for precise consumption forecasting. These findings inform effective energy management strategies across domains.

The advent of the Internet of Medical Things (IoMT) brings benefits and security challenges to the healthcare industry. Information about patients is becoming more vulnerable to cyber threats as the frequency of attacks on the healthcare sector rises. The patient's private and personal information, regularly collected through IoT devices and processed through a cloud server, is compromised. It can degrade everyone's reputation and cause significant financial losses. Patients' information must be protected from unauthorized use and access. This access control uses Fully Homomorphic Encryption (FHE) for this issue. This enhances information security schemes in healthcare systems by combining FHE and Blockchain technology. Our results show that the FHE and blockchain can be employed on low-power IoMT devices to provide varying levels of health data security. FHE allows calculating encrypted data without decryption and data protection, resulting in a safe and trustworthy system. The encrypted data can be handled and analyzed by authorized users without disclosing the file's contents, respecting the patient's privacy. This technique provides a secure and open environment for maintaining and transferring sensitive patient medical information.

The future of highway travel is being reshaped by autonomous vehicles (AVs). This microscopic study, conducted along a 9-kilometer highway in Ankara, Turkey, explores the dynamic relationship between AVs and travel time, as well as vehicle delay. Analyzing 17 scenarios with varying AV penetration rates (ranging from 25% to 100%) and diverse AV behaviors (cautious, normal, aggressive, and mixed) uncovered intriguing patterns. Cautious AVs, while promoting safety, introduced slightly slower travel times. In contrast, aggressive AVs prioritized efficiency and reduced travel times, striking a delicate balance between speed and safety. The introduction of mixed AV fleets demonstrated an exciting equilibrium, delivering competitive travel times and mitigating delays. Most notably, the presence of AVs in all configurations exhibited the potential to relieve congestion and enhance overall traffic flow. The findings offer a compelling microscopic perspective on the transformative potential of AVs in shaping the future of highway transportation. Understanding the complex dynamics of travel time and delay is critical for informed policy decisions and the evolution of urban mobility as autonomous vehicles (AVs) continue to improve.

Resource consumption is increasing globally. Global resource consumption has surged by over 65%, contributing to 70% of greenhouse gases (GHGs), raising concerns for future generations. Circular economy offers a strategic approach that perpetuates materials and regenerates nature, particularly with key materials, to potentially reduce GHGs by 40% by 2050. However, circular economy implementation faces challenges, notably in tracking material lifecycles. Due to the complexity of tracking material lifecycles, technology is needed, most notably, blockchain. This paper will review blockchain applications in the circular economy through existing literature. A total of 27 publications were reviewed where future research of the publications were also revealed.

Municipal solid waste (MSW) generation forecasting serves as the basis for future waste management strategic plans. However, rapid socio-economic and environmental changes lead to fluctuations in the MSW generation patterns. This is particularly exacerbated during unforeseen events, such as the COVID-19 pandemic. This study investigated the effect of COVID-19 and mobility conditions on the behavior of MSW generation predictors. Multiple artificial intelligence models were employed, including generalized linear models decision trees (DTs), support vector machines, deep learning, random forests (RFs), and gradient-boosted trees (GBTs). The models were tested around a 5-year historical MSW generation, incorporated with various socio-economic, seasonal, and climatic factors, in Abu Dhabi, United Arab Emirates. Multiple data split ratios were examined: 1) 60/40, 70/30, and 49/51 to simulate the COVID-19 pandemic. The feature importance analysis indicated that socio-economic had the highest impact on generation rates, with a Pearson correlation coefficient of 0.30-0.71, followed by seasonal and climatic factors, with 0.12-0.40 and 0.15, respectively. The forecasting results indicated that training the models with 70% of the dataset resulted in higher accuracy compared to using a 60% training split. GBTs and RFs demonstrated superior performance, with root mean squared error (RMSE) of 0.029 and 0.031, respectively. On the other hand, DT achieved the lowest accuracy of 0.057 RMSE. The analysis revealed a significant improvement in the models' predictive capabilities upon the inclusion of COVID-19 and mobility conditions by 17-41%. The findings of this study emphatically highlight the importance of incorporating contextual variables into MSW generation prediction methodologies, affirming their significant impact on enhancing forecasting accuracy in rapidly changing environments.

Parkinson's disease, a persistent and advancing neurodegenerative condition, impacts the CNS (central nervous system) and the mobility of body regions governed by neural pathways. During the initial phase of Parkinson's disease, individuals may experience early symptoms such as tremors in only one hand, stiffness, or decreased movement speed. Additionally, they may exhibit possible indications, including reduced facial expression, lack of arm swinging while walking, and a potential soft or slurred speech. The symptoms of this disease deteriorate as the condition advances, and the challenge lies in the fact that these symptoms vary from person to person and may often escape detection. Blood tests and physical scans do not provide a strong indication of the disease, and there are a lot of challenges in the data collection and processing of MRI images. However, speech changes in the early stage, also known as dysphonia, a bio-marker of the disease. This paper uses eight supervised machine learning models combined with four feature selection techniques. These models have been built considering their accuracy and F1 score, as medical analysis is susceptible to false positives and negatives in the classification process. A performance comparison of all the models has been analyzed and presented in this paper, and a study has been done. Support Vector Machine (SVM) is chosen as a final classification model as it gives a 97% F1 score and 93% AUC score with eight features, the best results among all combinations of classifiers and optimization techniques.

Smart surveillance has become an integral part of smart cities and urban living. Recently, unmanned aerial vehicle (UAV) or drone-based surveillance has become popular in smart cities for monitoring suspicious activities as well as for managing transport and emergency services. As the traditional surveillance systems operated by humans are constrained by human cognitive abilities and attention span, therefore a noticeable deterioration in object recognition can be witnessed. It is evident that human error can reach as high as 95% after about 22 minutes, which is significant for any catastrophic consequences. Additionally, automated surveillance systems, though effective, are demanding in terms of computing and memory resources. These computational and memory complexities lead to substantial energy consumption, making them inappropriate for remote surveillance, such as drone-based surveillance. In response to these challenges, we have engineered an adaptive system leveraging the deep learning architecture YOLO V7 in various quantization units. This has effectively reduced power usage by 1.6 times. This innovative system achieves a 22% decrease in graphics processing unit (GPU) consumption and reduces memory requirements by 29%, resulting in a surveillance system that outperforms traditional methods by being 1.43 times faster. This contributes to a more efficient solution for remote surveillance needs. This newly proposed system, with its significant reduction in power consumption, introduces a sustainable and robust surveillance alternative that retains its performance even in adverse conditions. The considerable speed boost facilitates a quick response, thereby making it a strong asset for maintaining security and swiftly addressing potential threats. Furthermore, we have evaluated this model's effectiveness for detecting suspicious objects, such as weapons, with an impressive accuracy rate of 93%. This makes it not only a more energy-efficient option but also an incredibly precise tool for identifying potential security threats.

The growth of Artificial Intelligence (AI) in recent years has transformed problem-solving across a wide range of areas. This study addresses the need for automated annotation solutions in the context of video analysis, specifically through the utilization of YOLOv8 models. There's an increasing demand for an accurate and efficient data annotation tool. The time-consuming and error-prone nature of manual annotation requires an in-depth examination of AI-based solutions. This study aims to evaluate the performance of the YOLOv5 and YOLOv8 models specifically the nano and medium versions in automatic video annotation. All the models are examined for their balance between speed and accuracy, to test their applicability in real-time. The evaluation process involved various key performance metrics including mAP, precision, recall, and training time per epoch were evaluated after the models were trained on annotated datasets consisting of 300 images. YOLOv8n outperforms the other models with a mAP of 90.9%, 88.5% precision, and 89.9% recall. This model achieves the best possible balance of speed and accuracy, making it a promising option for real-world applications. Explainable AI (XAI) approaches, particularly Grad-CAM, were used to improve interpretability. Visualizations show the model's decision-making process, which promotes transparency and confidence. The discussion explores the implications of model selection, limitations, and outlines techniques for continued refinement.

This study investigates how blockchain and Federated Learning (FL) may work together to create a patientcentered, long-lasting healthcare ecosystem. The paper presents an innovative method for improving data privacy, security, and interoperability within healthcare systems by utilizing the distributed Medical Internet of Things (MIoT). The next generation of healthcare services will be revolutionized by the synergistic combination of FL and blockchain, enabling a decentralized framework for enhanced patient care and data-driven decisionmaking. Centralized Machine Learning (ML)-based methods lack security of patient's sensitive medical information as all information of patient are trained in central server. To tackle with this problem, this paper offers amalgamation of FL and blockchain technology which enables training of ML models collectively by making use of distributed MIoT data, enhances the accuracy and also guarantees privacy of patient's personal medical record. To preserve privacy over patient's personal medical record, blockchain technology is used. The proposed approach offers a lot of benefits: (1) To predict disease more perfectly and effectively, this paper utilizes AdaBoost, Decision Tree (DT), Extra Tree Classifier (ETC), Linear Discriminant Analysis (LDA) and FL algorithm. (2) Each algorithm is tested against different data splitting, different cross validation and cross validation with feature selection techniques, hyperparameter tuning for model optimization and finding the best model. (3) The proposed methodology ensures collective learning among several healthcare organizations without sending patient's private medical records and thus safeguarding privacy. (4) Accuracy, Markedness (MK), False Omission Rate (FOR), Negative Likelihood Ratio (NLR), Prevalence Threshold (PT) are calculated to choose the best algorithm and also to show how well the FL algorithm works. (5) The paper is unique as the best algorithm in terms of different model evaluation metric is first find out and then that algorithm is fed to FL. The results indicate higher accuracy and compared to the modern techniques.

Internet of things (IoT) brings unparalleled modifications to all contexts of our lives, as they may be knowledgeable via way of means of smart devices and real-time data. Among the various, IoT software settings, e-government appears to be one which may be significantly benefited via way of means of using IoT, remodeling and augmenting public offerings. This paper targets to make contributions to a higher information of the way IoT may be leveraged to enhance e-government. IoT adoption in e-government encompasses numerous demanding situations of technical in addition to organizational, political, and criminal nature, which ought to be addressed for growing efficient applications. With the software of IoT in e-government being at an early stage, it's miles vital to analyze those demanding situations and the methods they might be tackled .E- government seems to be one of the many IoT software settings that could advantage considerably. There are converting and augmenting public offerings via using IoT.

Hypertension or high blood pressure is the major cause of several health complications and increases the chances of heart disorders, stroke, and sometimes death. Therefore, early detection of hypertension will save many lives. This paper discusses some of the most preferred machine learning techniques namely: Convolutional Neural Network (CNN), Artificial Neural Network (ANN), Naive Bayes, Support vector machine (SVM), Random Forest, multi-layer perceptron (MLP), and Logistic Regression which helps in the pre-diagnosis of hypertension. This study will help the researchers in finding out the datasets and understanding the performances of the machine learning algorithms in terms of accuracy, sensitivity, and specificity.

Urbanization has made waste management a major environmental and health issue. Growing populations and urban landscapes have overwhelmed traditional garbage management systems. Urbanization and fast city growth have made garbage management harder. Unethical trash disposal pollutes the environment and harms people. Traditional waste management systems typically lack efficiency and communication, delaying waste collection and resource allocation. This paper offer a Smart City Waste-Management Application that integrates hardware and software to solve this problem. An ESP_32 microcontroller and ultrasonic sensors assess waste bin fill levels in real time for a Firebase database. QR-coded smart bins update this data with map-based locations. A smart bin contains an update information through its embedded QR codes which display the map of its location. The software component is categorized into three primary sections: They are composed of Administrator, Driver and User, dedicated to particular activities. The Administrator manages system operations and is able to publish advisory publications. This is where the driver section assists the drivers with performing their tasks on collecting the wastes efficiently. The user segment enables people to sign up for the service, where they can interact with other participants through the chat page. The app also comes with a ``Map'' page which monitors driver movements as well as location of waste containers thereby saving on travel time. The activity page stands out to be of critical importance with a section on drivers' activities and another one for following up on users' payments. Moreover, it has a subscription payment facility where subscribers can log in and access the application. Therefore, this Smart City Waste Management model integrates various elements including the ESP_32 and the ultrasonic device coupled with diverse software tools. The objective of this project includes straightening up these waste collection procedures, making them better for users, as well as effective waste management within the city.

Each year around 40 to 80 thousand deaths occur due to misdiagnosis in US. Inaccurate diagnosis remains the first cause of medical errors which cost around 1.8 billion $. This paper proposes a diagnosing system using three models: decision tree, random forest, and gradient boosting algorithms for 41 disease detections using two datasets. The results from the conducted analysis using the first dataset suggest that the gradient boosting model demonstrates the best performance achieving an accuracy of 93.9% and 6.1% misclassification rates. For the second dataset, random forest outperforms the other models giving the highest accuracy of 99.59% and the lowest misclassification rate of 0.4%. The k-fold cross-validation method compared with the splitting data accuracy improved the model's performance. The t-statistical test shows that the proposed models were statistically significant.

The implementation of Software-Defined Networking (SDN) in Flying Ad-hoc Networks (FANETs), known as SDN-FANET, has gained significant traction in the past decade. This is particularly due to its eccentric capabilities and features which effectively enhance the operation and performance of Unmanned Aerial Vehicles (UAVs) in various applications. SDN-FANETs introduces adaptability and programmability to the network by centralizing its control. A single entity known as the SDN controller, carries out the network control's management and operation. However, this centralized approach introduces a significant risk: a Single Point of Failure (SPoF) associated with the SDN controller. The failure of the controller can lead to severe network disruptions. Therefore, to tackle this challenge, recent research has focused on decentralizing the SDN controller. This survey presents, compares, and analyzes the latest work done in mitigating the centralization issue. Various approaches to decentralization, such as multi-controller architectures, distributed control, and hybrid models, are discussed based on their effectiveness in improving network scalability, reliability, energy efficiency, routing, security, and cost of implementation. In addition, we categorized these schemes in terms of the main technologies utilized: Cloud & Edge, Blockchain, a hybrid of both, and miscellaneous ones. Our work also highlights important findings from the discussed solutions, and open research issues in the context of SDNFANETs along with some possible future directions. By identifying these research gaps, our survey contributes to the understanding of future research directions in the field of SDN-FANETs.

Cybersecurity has become a vital concern these days, which is why many approaches to data protection are constantly being invented. However, complex systems are difficult to manage, and manual data security is particularly ineffective in these cases. Therefore, Security Orchestration and Automation Response (SOAR) is one of the leading solutions in the field of data security. The following paper provides an account of this technology in the context of its application with predictive risk control systems by using machine learning.

In this paper, a method for protecting the privacy of private health information called information-fused blockchainbased Federated Machine Learning-based disease forecasting utilizing Intrusion Detection System (IDS) on Internet of Health Things (IoHT) is presented. Privacy and secrecy are the main IoHT problems because of the size and extensive implementation of IoHT devices. The development of ``Smart Healthcare'' is a result of Machine Learning (ML) strategy, which has greatly improved the capacities and amenities of healthcare 5.0. As a consequence, patients' quality of life will increase, and stress levels and medical costs will decrease. Among the data-driven technological characteristics made feasible by the IoHT is smart, collaborative healthcare. However, combining healthcare information into a single storage site is necessary to create a potent ML model, which creates security, regulatory and ownership difficulties. By propagating a universal framework for learning via a centralized aggregate service, Federated Learning (FL) addresses these problems. The local strategy maintains control over the patient data, preserving its privacy and confidentiality. The following advantages of the recommended course of action: (1) This paper carefully examines the use of blockchain-based FL for healthcare 5.0. (2) The technique aims to provide a secure healthcare surveillance system for healthcare 5.0 by using an IDS to detect any malicious behavior in the healthcare industry. (3) The proposed system performs more effectively for forecasting intrusion detection than the healthcare system linked to the FL approach suggested in the literature.

This study explores the integration of deep learning and computer vision techniques for contactless biometric identification, specifically focusing on wrist bracelet lines. In an era where biometric identification plays a pivotal role in various sectors, including personal identity, mobile devices, and smart gadgets, this research addresses the heightened demand for efficient contactless solutions, amplified by the COVID-19 pandemic. While conventional methods such as finger knuckle, face recognition, and fingerprint analysis have been prevalent, this research introduces an innovative approach that capitalizes on the distinct and enduring patterns observed in wrist bracelet lines. The proposed system employs advanced deep learning methodologies, notably the YOLO (You Only Look Once) model for precise wrist detection. To facilitate image capture and transmission for identification, a combination of an Arduino Uno and an ESP32 CAM module is employed. The study highlights the system's real-time recognition capabilities through comprehensive results, emphasizing its potential for secure, efficient, and contactless biometric identification in Smart city applications.

The widespread use of Android smartphones in daily life can be attributed to the extensive prevalence stemming from the Android OS and the availability of opensource applications. People have become accustomed to performing various tasks, such as online banking transactions, scheduling appointments, attending virtual meetings and classes, and online shopping, on their mobile devices instead of computers. However, the significant use of mobile phones also presents opportunities for attackers to access sensitive and confidential information without the user's knowledge. Intruders often use malicious applications found on the Google Play Store and other third-party app markets. The increased volume of Android malware in recent years has severely threatened human privacy and security. In this paper, we present a novel solution that utilizes a stacking ensemble machine learning approach to detect whether an application is malware or benign, which achieved 98% accuracy and F1-Score, and to identify the malware category with 96% accuracy and F1-Score. We conducted the experiment on the CCCS-CIC-AndMal-2020 dataset, which comprises 27,125 samples. Static features such as permissions, meta-data, and intents were analyzed, and the performance of standard machine learning algorithms was compared to our proposed model. Our results demonstrate that the stacked ensemble model shows promising results for malware detection (binary classification) and malware type classification (multi-class classification).

Due to its broad usage and predominance, TCP (Transmission Control Protocol) serves as the core of all Internet traffic. TCP is widely used because of its reliable connection-oriented architecture and capacity to manage congestion using different congestion control algorithms across a range of network loads and circumstances. The three phases of these congestion control algorithms-Slow Start, Congestion Avoidance, and Congestion Detection (Fast Recovery and Fast Retransmit)-are generally how they work. However, different TCP variants, such as Reno, New Reno, Tahoe, and more recent versions like FAST TCP, Cubic TCP, and Elastic TCP, implement these phases quite differently. Ever since these protocols were developed, the Internet environment has changed significantly, and they now have to deal with diverse set of challenges like high bandwidth demand problem and the requirement of fairness among various TCP flows. In this paper, we will review the most recent literature considering TCP variants for Internet of Things (IoT). Furthermore, we will analyze these techniques based on various factors, such as technology used, evaluation methodology and metrics, nodes type and mobility.